As is well known and understood, bar codes are used by the United States Postal Service and other mail carriers in an attempt to speed up the processing of mail delivery and to reduce labor expense. As is also well known and understood, these bar codes are typically sprayed onto the mail documentation by such expensive equipment as optical character readers which first read the addresses electronically, and then converts the address into machine readable codes for further processing.
However, and as has been widely publicized, the equipment employed in such sophisticated optical-scanning technology and computer-based artificial intelligence systems experience great difficulty in deciphering street names, numbers and Zip-Codes written in pencil or pen. As has also been highly publicized, this is particularly a problem where the information is handwritten. In such instances--amounting to upwards of 30-40 percent of United States Postal Service delivery--, in order to process the mail the bar code is affixed to speed delivery; obviously, this represents both a tedious and costly task.
In an attempt to combat this, much effort has been devoted to analyzing thousands of handwritten and printing styles, to try to determine quirks and commonalities that computers can be taught to identify and interpret. As has been reported in the print media, researchers have demonstrated experimental systems that can discern enough information from a handwritten envelope to meet the needs of automated letter-sorting equipment--but these equipments have been determined to take a far longer period of time to read the address than the Postal Services desired speed of some 13 pieces of mail per second for its automated sorting machines.
As has also been widely advertised, it is the present desire of the United States Postal Service to have bar codes on all mail by the year 1995, so as to have the coding mail system completely operative by then. In view of that, much sophisticated information-technology and government-contracting corporations are proposing systems to obtain address-recognition contracts to replace or supplement the automated scanning and sorting equipment currently being used. When one factors in that such "handwritten recognition" could also be useful in the banking, credit card and/or other industries, in which large volumes of handwritten documents jam-up computer processing, it is not too hard to understand all the engineering time and effort being invested.
However, equally apparent is the lack of attention which the Postal Service has devoted to easing this problem at the "sending end," where the documents are being addressed. While admittedly, the present Postal operation gives a premium of 2 cents or more for every piece of first-class mail that is already addressed with a bar code, that fact has not been very-well publicized. In fact, by-and-large, the Postal Service has come to a seeming conclusion that nothing it can do can bring the "sending public", into the equation of facilitating the mail delivery by typewriting the envelope so that its address can be computer recognized, or by having envelopes already pre-printed with bar code destinations, as have been employed by public utility companies, credit card companies and insurance companies in processing their already addressed coupons to be returned by a consumer along with its monthly or periodic payment.